Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Charlotte Godard, Corentin Guérinot, Valentin Marcon, Mohamed El Beheiry, Jean-Baptiste Masson (Institut Pasteur, 25-28 rue du Dr Roux, 75015 Paris, FRANCE)
Recently, advances in volumetric imaging have enabled the acquisition of high resolution images, leading to an increase in data size and complexity of structures. Visualization and analysis of complex 3D data remain a challenge as actions like segmentation and selection are tedious on a 2D screen. We introduced a software, DIVA, for analyzing and visualizing any type of 3D images in virtual reality (VR) without pre-treatment. Based on a dual desktop-VR interface, it allows a complete immersion inside the volumetric data, performs rendering through optimized ray-tracing and enables live modification of data appearance with a user-friendly transfer function interface. Yet, while VR shows the promise to strongly enhance visualization, what is really needed is a mix of visualization and treatment. This creates a computational challenge since VR is very demanding. We introduce a new software, DIVA cloud, combining the DIVA VR visualization with cloud computing allowing both visualization and live data treatment. We demonstrate the possibilities of live inference while interacting with them in a one-shot learning segmentation task i.e. segmentation without training on a database.